https://github.com/rsanchezgarc/bipspi
Partner specific prediction of protein binding sites
https://github.com/rsanchezgarc/bipspi
bioinformatics machine-learning protein-protein-interaction proteins structural-bioinformatics
Last synced: 6 months ago
JSON representation
Partner specific prediction of protein binding sites
- Host: GitHub
- URL: https://github.com/rsanchezgarc/bipspi
- Owner: rsanchezgarc
- License: apache-2.0
- Created: 2018-05-24T14:43:18.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2023-03-13T10:44:04.000Z (over 2 years ago)
- Last Synced: 2025-04-10T23:08:09.267Z (6 months ago)
- Topics: bioinformatics, machine-learning, protein-protein-interaction, proteins, structural-bioinformatics
- Language: Python
- Size: 950 KB
- Stars: 8
- Watchers: 0
- Forks: 9
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# BIPSPI+: Enhancing partner-specific binding site prediction with better data
[BIPSPI](https://bipspi.cnb.csic.es/) (xgBoost Interface Prediction of Specific Partner Interactions) is a partner specific
binding site predictor that employs the XGBoost algorithm. This new version, BIPSPI+, has been trained on
newer and larger datasets compiled from the PDB, severely improving its performance.
BIPSPI+ can be employed to predict partner-specific binding sites given two atomic models,
two sequences, or one sequence and one structure.
BIPSPI+ is distributed as a Docker image and as a GitHub repository. Installation is only required in
the latter case. Complete guide can be found in:
- [DOCKER image](docs/docker_help.md)
- [GitHub repository](docs/repo_help.md)
### New features
- Homo-complexes and hetero-complex specific models, boosting performance especially when predicting homo-complexes.
- Sequence vs structure mode. Original version could only be executed with two sequences or two structures. Now one sequence and one structure could be used as input as well.
- Atomatic correlated mutations (optional)